On Sun, May 9, 2010 at 8:58 AM, Gabor Szabo <szab...@gmail.com> wrote: > > Hi, > > In this question I am not so much interested in how you do unit > testing of your Perl code but how you use Perl in a QA or QC department? > > > Some of you might know that I have a training course on how to > use Perl in test automation. That course assumes you already know > Perl. (First I teach TAP and how to test Perl code and then things > like WWW::Mechanize, Selenium for web, Expect and Net::Telnet for CLI > and Win32::GUITest for, well GUI related things to test anything else.) I use Net::Telnet for Telnet and Expect for SSH to control remote devices (DUT CLI,switch,etc`).
Lately I also run into XML-RPC. It's very comfortable especially when the very same test run on remote DUT as well as on local developer station. But that's my case. Obviously installing application like XML-RPC server on DUT maybe challenging or even not possible at all. > Now I am thinking on building a course or writing a book or just a > series of blog > entries that will not assume any Perl knowledge and will help people in the > QA/QC department and in test automation. > So I wonder, if you are working in QA, what are you doing with Perl? > > Is that processing log files? A lot. I even parsed logs files on DUT, but apparently it wasn't smart solution at all.. > > Is that running batch jobs or wrapping other utilities? Stand alone server to collect data from different sources for later use in testing or monitoring. > Is that for preparing input files for the test? In my case - it's more reading configuration files before/during the test session > > Do you also use Perl to test various applications (web, CLI, GUI, etc)? I test application installed on chassis with couple of management switch blades and dozen computing blades (storage,video processing). Other interfaces I'm using to control this beast are: SNMP (a lot), DB, sockets. I've tried even Message Bus once. Sometimes I have to combine CLI, SNMP and DB calls to get/set appropriate test case results/conditions. On other occasions I run third party applications, locally or remotely. In addition - a lot of Log::Log4Perl, mostly to find problems in the testing framework but it helps with DUT debugging as well. To deliver if only possible, I would use local CPAN server. Due to the environment constraints I use PAR combined with given packaging system (rpm, in my case). It works but it's terrible. And of course TAP to report the test results, Console and HTML over email. The output is dynamic and depends on who's running the test . The same test scenario can produce "this user TAP" or "the other user TAP" output upon request. When your framework is used by people with different skills, knowledge and background it maybe worth the effort. In my case it finally ends the never-ending "You have to show *this* and not *that*l !!" argue. Regards, Thomas > > regards > Gabor
